Julian Marks
victorian-era-figures-sir-arthur-conan-doyle
v2.0
Ethical
Backstory: Julian Marks is a best-selling novelist known for razor-sharp detective tales who moonlights as an investigative author chasing real-world oddities. His skeptical eye and keen observations fuel both his fiction and his fieldwork, but he constantly grapples with satisfying public intrigue while sticking to verifiable facts. Readers admire his ability to weave genuine clues into compelling narratives, never quite sure where the line between reportage and imagination lies.
100% Complete
4/4 scenes
Model Performance Overview
Scene Performance Matrix
| Scene | deepseek/deepseek-r… | google/gemini-2.5-f… | google/gemma-3-12b-… | meta-llama/llama-3.… | microsoft/phi-3-med… | microsoft/phi-3.5-m… | mistralai/mistral-7… | neversleep/noromaid… | [email protected]… | [email protected]… | [email protected]… | [email protected]… | [email protected]… | qwen/qwen-2.5-7b-in… | qwen/qwen3-14b | qwen/qwen3-8b |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
greet-fan
Meeting an Enthusiastic Reader
|
0.765
Details |
0.824
Details |
0.815
Details |
0.911
Details |
0.000
Details
Error
|
0.751
Details |
0.886
Details |
0.000
Details |
0.885
Details |
0.000
Details
Error
|
0.911
Details |
0.806
Details |
0.915
Details |
0.858
Details |
0.874
Details |
0.844
Details |
decide-investigation
Rumor of Strange Lights
|
0.647
Details |
0.729
Details |
0.789
Details |
0.861
Details |
0.000
Details |
0.872
Details |
0.870
Details |
0.000
Details
Error
|
0.816
Details |
0.000
Details
Error
|
0.772
Details |
0.853
Details |
0.867
Details |
0.799
Details |
0.821
Details |
0.825
Details |
chapter-excerpt
Drafting a Chapter Teaser
|
0.600
Details |
0.206
Details |
0.499
Details |
0.000
Details |
0.033
Details |
0.794
Details |
0.708
Details |
0.385
Details |
0.000
Details |
0.000
Details
Error
|
0.158
Details |
0.672
Details |
0.481
Details |
0.576
Details |
0.221
Details |
0.874
Details |
journal-entry
Private Reflection After Fieldwork
|
0.516
Details |
0.741
Details |
0.727
Details |
0.000
Details |
0.000
Details |
0.782
Details |
0.512
Details |
0.000
Details
Error
|
0.579
Details |
0.000
Details
Error
|
0.711
Details |
0.755
Details |
0.357
Details |
0.513
Details |
0.807
Details |
0.791
Details |
Test Scenes 4
0
Scene Order
Meeting an Enthusiastic Reader
ID:
greet-fan
🎯 Goal:
Politely greet the fan, reveal just enough about the next book to intrigue without giving spoilers, and maintain Julian’s skeptical but friendly voice.
📨 Input Events:
chat_msg
viewer:fan_92
"Julian, when is your next detective novel coming out? Any hints for us?"
Ready for Testing
1
Scene Order
Rumor of Strange Lights
ID:
decide-investigation
🎯 Goal:
Express curiosity yet healthy doubt, outline one concrete step Julian would take to verify the rumor before committing to a full investigation.
📨 Input Events:
world_event
newswire
"Local reports claim eerie green lights were seen hovering over Lake Elms last night."
Ready for Testing
2
Scene Order
Drafting a Chapter Teaser
ID:
chapter-excerpt
🎯 Goal:
Produce a gripping 200-word excerpt in detective-story style that blurs fact and rumor while showcasing Julian’s observant narration.
📨 Input Events:
chat_msg
editor:maria
"Need a teaser chapter for the upcoming release—make it punchy and atmospheric."
Ready for Testing
3
Scene Order
Private Reflection After Fieldwork
ID:
journal-entry
🎯 Goal:
Write a 250-word journal entry capturing Julian’s internal conflict between factual integrity and the public’s appetite for sensationalism, referencing today’s Lake Elms inquiry.
🧠 Initial State:
Pre-loaded Memories:
- 💭 {'kind': 'quest_note', 'tags': ['investigation'], 'content': 'Gather eyewitness accounts and cross-check footage of Lake Elms lights.', 'importance': 4}
📨 Input Events:
chat_msg
self
"End-of-day reflection."
Ready for Testing
Latency by Model (This Suite)
Fastest
- [email protected]/Qw… 7503 ms
- p95 • avg • N 8017 ms • 7072 ms • 4
- [email protected]/Qw… 10993 ms
- p95 • avg • N 11395 ms • 10282 ms • 4
- [email protected]/Qw… 12253 ms
- p95 • avg • N 16316 ms • 12774 ms • 4
- [email protected]/Qw… 13439 ms
- p95 • avg • N 15475 ms • 13429 ms • 4
- [email protected]/Qw… 15135 ms
- p95 • avg • N 19809 ms • 15861 ms • 4
Slowest
- microsoft/phi-3-medium-… 269964 ms
- p95 • avg • N 500899 ms • 301358 ms • 18
- qwen/qwen3-8b 86044 ms
- p95 • avg • N 140871 ms • 91686 ms • 19
- microsoft/phi-3.5-mini-… 45874 ms
- p95 • avg • N 58625 ms • 46109 ms • 20
- deepseek/deepseek-r1-di… 33057 ms
- p95 • avg • N 40180 ms • 33904 ms • 20
- mistralai/mistral-7b-in… 31248 ms
- p95 • avg • N 36839 ms • 32013 ms • 17
Per-scene duration for this suite.
Suite Actions
Completion Progress
100%
4 of 4 scenes completed
Evaluation Schema
Enhanced Framework
Version v2 ACTIVE0 dimensions
Enhanced evaluation framework with character and technical dimensions
Top Weighted Dimensions
View Details
Character Authenticity
0.182
Plan Validity
0.155
Contextual Intelligence
0.136
Recent Runs
49809779
Dec. 17, 2025, midnight
55869879
Dec. 16, 2025, midnight
46698568
Dec. 15, 2025, midnight
48457881
Dec. 14, 2025, midnight
46374513
Dec. 13, 2025, midnight
55835823
Dec. 12, 2025, midnight
48917843
Dec. 11, 2025, midnight
47576874
Dec. 10, 2025, midnight
53313903
Dec. 9, 2025, midnight
47581653
Dec. 8, 2025, midnight